Re: [PATCH net-next v2 1/2] u64_stats: Allow u64_stats_add() to accept s64

[David Laight <david.laight.linux@gmail.com>]

On Mon, 26 Jan 2026 06:07:49 +0800
David Yang <mmyangfl@gmail.com> wrote:

> The current u64_stats_add() accepts only unsigned long, which makes no
> sense as u64_stats_set() accepts u64 already, and causes unnecessary
> narrowing on 32-bit architectures.

Isn't is a chicken/egg?
A value over 2^31 could just be a very large value that needs to be added.
Where the add is being done 'just in time' so avoid overflow.

	David

> 
> Moreover, since there's no u64_stats_sub(), users may be tempted to
> reuse it for decrements, which already happened on nft_counter_reset()
> in net/netfilter/nft_counter.c .
> 
> Change to s64 to allow both positive and negative values.
> 
> Signed-off-by: David Yang <mmyangfl@gmail.com>
> ---
>  include/linux/u64_stats_sync.h | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/include/linux/u64_stats_sync.h b/include/linux/u64_stats_sync.h
> index 457879938fc1..e77f2b30bac7 100644
> --- a/include/linux/u64_stats_sync.h
> +++ b/include/linux/u64_stats_sync.h
> @@ -84,7 +84,7 @@ static inline void u64_stats_set(u64_stats_t *p, u64 val)
>  	local64_set(&p->v, val);
>  }
>  
> -static inline void u64_stats_add(u64_stats_t *p, unsigned long val)
> +static inline void u64_stats_add(u64_stats_t *p, s64 val)
>  {
>  	local64_add(val, &p->v);
>  }
> @@ -125,7 +125,7 @@ static inline void u64_stats_set(u64_stats_t *p, u64 val)
>  	p->v = val;
>  }
>  
> -static inline void u64_stats_add(u64_stats_t *p, unsigned long val)
> +static inline void u64_stats_add(u64_stats_t *p, s64 val)
>  {
>  	p->v += val;
>  }